It would be nice if everyone was equally brilliant, but in reality, people differ in their cognitive ability. Or should that be “cognitive abilities?” If you call me smart, is there one brain network underlying my performance in a wide variety of mental tasks? Or do I happen to be good at a variety of tasks, each with their own underlying brain networks? These questions have been debated for decades without a definitive answer, but important progress was made in one of the largest-ever studies of intelligence.
The study, called Fractionating Human Intelligence (yes, apparently “fractionating” is a real word), published in high-profile journal Neuron, has three properties that set it apart from past research. First, it used neuroimaging to more directly examine the brain networks involved in cognitive function. Second, it combined neuroimaging with behavioural, computer-based online testing from Cambridge Brain Sciences instead of the pen-and-paper tests that have been used in—and some say holding back—research for decades. And third, because of the advantages of online testing, its scale was massive; 44,600 participants took the cognitive tests.
Traditional intelligence studies looked for patterns in behavioural tasks, and often came to the conclusion that all tests are correlated, so there must be a single “g” factor behind all variation in intelligence. However, there might be other explanations, and this study narrowed down the possibilities by observing which brain networks are activated during cognitive behavioral tests.
The study’s complex and controversial results cannot be fully summarized in a single post, but let’s look at some key findings.
This figure shows an analysis of brain activation during twelve Cambridge Brain Sciences tests. A factor analysis technique pulled out two distinct brain networks. One network was activated by tests that required maintaining information in short-term working memory (the red bits). The other network was activated by tests that required transforming information in the mind according to logical rules, or reasoning (the blue bits).
The researchers took it a step further. They analyzed individual differences in performance on the same cognitive tests in a larger population, and looked for patterns in this behavioural data. There was one pattern that stood out: tests that relied on working memory clumped together, and tests that relied on reasoning clumped together. Oh, hey, wow, that’s the same pattern as in the brain imaging data.
Further analysis revealed a third factor as well: tests that required verbal skills. The imaging data confirmed that this factor, too, has its own separate brain network to support it.
Intelligence is More Than One Thing
So why did past intelligence studies often conclude that there is only one “g” factor? Further analyses and simulations revealed that there are correlations between various tests, but they can be fully accounted for by tests corecruiting distinct brain networks. In other words, most tests rely on more than one brain network, so performance is correlated between tests, but those underlying networks are completely separate.
“Human intelligence is most parsimoniously conceived of as an emergent property of multiple specialized brain systems, each of which has its own capacity.”— Hampshire, Highfield, Parkin, & Owen, Fractionating Human Intelligence
What does it mean to be smart? This evidence suggests there are at least three different ways to be smart: having well-functioning working memory, having well-functioning reasoning, and having good verbal skills. Different people will be better or worse at different tests, depending on which brain networks the tests require.
People don’t differ in cognitive ability. They differ in cognitive abilities.
The research above is one of more than 300 peer-reviewed studies that used tests in the CBS Trials platform—a leading web-based service for the assessment of cognitive function. Visit CBS Trials for more information and to try it in your own research.
Source: Fractionating Human Intelligence